自动驾驶车辆道路跟驰与状态一致性控制

Path following and status consistency control of autonomous vehicles

针对含有未知输入和外部干扰的非线性自动驾驶车辆时变跟驰队列系统,研究系统部分状态可测情况下的车辆道路跟驰和状态一致性控制问题.基于车辆跟驰和二自由度动力学模型,得到含有外部干扰和未知输入的离散化状态方程,利用前导跟驰特性,得到自动驾驶车辆跟驰队列系统;利用比例积分状态观测器解决系统部分状态不可测问题,提出一种基于观测器实现含有未知输入和外部干扰的自动驾驶车辆状态一致性控制协议;将观测器估计效果和一致性控制问题转化为误差系统的稳定性问题,由此构造Lyapunov-Krasovkii函数,利用离散系统稳定性理论推导出一个充分条件;通过求解线性矩阵不等式(linear matrix inequality, LMI)得到跟驰系统的增益矩阵和参数矩阵,利用H_∞性能指标分析系统鲁棒性.仿真结果表明:所设计观测器能够估计未知输入、外部干扰和系统状态,并且基于观测器设计能够使自动驾驶车辆道路跟驰和状态达到一致.

The problem of path following and state consistency control in a nonlinear time-varying platooning system of autonomous vehicles with unknown inputs and external disturbances is investigated in this study.Specifically,the focus is on addressing the control challenges when only partial states of the system can be measured.First,using the vehicle-following and two-degree-of-freedom dynamics models,a discrete state equation is derived for the platooning system of autonomous vehicles,which considers external disturbances and unknown inputs.The leading-following characteristics are utilized to establish an autonomous vehicle platooning system.Then,the partial state unobservability problem of the system is addressed by employing a proportional-integral state observer.An observer-based approach is proposed to realize state consistency control for autonomous vehicle platooning systems.In addition,the estimation effectiveness of the observer and the problem of state consistency control are transformed into stability issues of the error system.This is realized by constructing a Lyapunov-Krasovskii function and deriving sufficient conditions using the discrete system stability theory.The robustness of the system is examined using the H∞performance index.Finally,the gain and parametric matrices of the platooning system are obtained by solving the linear matrix inequality(LMI).Simulation results verify the effectiveness of the observer-based design in allowing accurate path following and ensuring state consistency for autonomous vehicles with unknown inputs.